Words and photos by Matt Kelly
One of the biggest steps we took in our human evolution was the realization that we can exert control over the food we eat.
At some point, way back in the early dawn of civilization, a community realized that we didn’t simply have to rely on gathering and chasing down wild things. We could actually domesticate food and produce it within our communities. Agriculture gave us a little more certainty in the struggle to fill our bellies and stay strong.
For millennia we’ve continued to expand on this process, breeding crops for specific traits: better taste, higher nutritional content, more abundant yields, longer shelf life, greater resistance to pests. We’ve done this through a very basic process: plant our crop, grow it out, and then save seeds from the individual plants that show signs of the traits we want. Over and over, year after year, we’d repeat these steps until we ended up with traits that consistently showed up every time we grew that particular variety. We learned how to guide the natural evolutionary process of the crops in a direction that benefited us. Corn is a perfect example: over the course of centuries we selected for individual plants that eventually changed the grain from grass-like stalks into the big, fat ears we know today.
This way of altering our food – of making it less wild and more a part of human communities – is a slow process, often taking generations to get solid results. It’s also fraught with uncertainty; weather, disease, pests and even our own human whims can have unpredictable impacts. But it does work. And this breeding approach is based entirely on selection by phenotype: making choices about physical traits that we can observe with one or more of our five senses.
The set of observable characteristics of an individual resulting from the interaction of its genotype with the environment.
The genetic composition of an individual organism.
Of course, figuring out how to domesticate food isn’t the only thing humans have been doing all this time. We’ve also been making tools, often driven by our aspirations to be better farmers. In fact, we’ve gotten so good at making tools that humans no longer need to depend solely on phenotypes when it comes to crop breeding. We can now access the molecular information behind those traits and make selections based on genotypes – something we’ve never been able to do with our five senses alone.
In one of the many labs at Cornell University, a group of young researchers and lab techs are leveraging these new tools to make breeding crops more efficient and more effective than it’s ever been. We’re not talking about artificially manipulating the genetics of plants. Instead, we’re talking about using some of the latest and greatest genetic technologies in conjunction with the same breeding practices that are nearly as old as humankind itself.
We’re talking about marker-assisted selection.
“Today’s growers want a Delicata squash that is better than the Delicata they already have on the market,” says Michael Mazourek, assistant professor and plant breeder at the Cornell School of Integrative Plant Science. He opens the first set of doors into the greenhouse. “Growers want good yields, they want early yields, and they want a Delicata that stores a little longer. And everybody always wants sweeter taste.”
Mazourek opens another set of doors. “Of course, that’s a bit of a paradox with the Delicata. It’s so delicate you can eat the skin. But because it’s thin-skinned, it’s not very durable or long-lasting.”
Apparently, breeding the perfect crop isn’t easy.
Beyond those second doors, the greenhouse is filled with hundreds and hundreds of small squash plants in four-inch pots. Rachel Hultengren, a graduate student in Mazourek’s research group, is looking them over. “I plan on getting rid of three quarters of the plants in here," she says. "I’ve got about 450 right now; I plan to end up with about a hundred plants.”
She plans on getting rid of all those plants because they will likely be susceptible to powdery mildew, a common fungal disease that can be a serious problem for growers. Mazourek and Hultengren only want Delicatas that are resistant to the fungus. Sure, resistance to the pathogen is an excellent trait to have in a squash; but just as important, Mazourek and his team need to be certain the squash plants they’re working with don’t die off before they actually produce fruit that can be assessed for the other traits.
Like sweeter flavor.
All of this could be accomplished with traditional breeding methods. But even under the best modern circumstances, it would take years – maybe a decade – to accomplish. “What’s great about marker-assisted breeding is that it speeds up the process,” says Hultengren. “It allows us to devote more resources to a lot fewer plants. It means we spend a lot less time in the field trying to evaluate whether or not a squash has resistance. It allows us to focus only on plants that have the genetics that we want as parents for future lines.”
The process of using genetic markers to assist in breeding goes something like this: Hultengren will collect one tender little leaf from each plant and bring them all back to the lab. She’ll grind up the samples to get at the genetic information in the cells and then test those samples to see which of these hundreds of plants have the genetic variation for resistance. She knows exactly where to look for the variation because of known markers in the genome of the squash.
This saves immense time, energy, effort and resources that would normally be spent growing out a whole field of plants just to find the few that have the one trait the team is looking for.
So, how about looking at markers for sweetness and durability at the same time?
“We’re getting there,” says Mazourek. He's trying to set reasonable expectations. “Resistance to powdery mildew is one gene. It’s a yes or no thing. But sweetness in squash is much more complicated. There’s a dozen – if not hundreds – of genes all participating in how sweet the squash is going to be.”
Hundreds of genes mean hundred of markers to look for.
Kyle LaPlant sits in front of a double-wide set of computer screens in the team’s lab, intensely working his way through the genome of stringless peas. He’s another grad student on this research team. “This is a broad representation of markers across the genome for a population of stringless peas,” he says, pointing at the grid of numbers, letters and countless colored boxes.
Markers are like a street address. It’s a numbering system humans have assigned to specific spots on a genome to help guide us along the windy genetic road for a particular organism. At each marker is a pair of bases: biological compounds that are the smallest genetic building blocks. A specific sequence of these base pairs that are connected to a distinct physical trait is what we call a gene.
And genes – of course – determine whether a pea is stringless or a squash is resistant to powdery mildew.
“Sometimes it’s as simple as going through each marker and assessing whether or not it’s significantly associated with a specific phenotype,” says LaPlant. In other words: Do stringless peas always have the same base pair at a particular marker? That’s as simple as it can get if the trait is just associated with a single gene. “For more complex traits like height or yield, you would have to find many locations in the genome that significantly associated with that phenotype,” he says. Multiply the time it takes to identify a single marker by however many genes are involved.
Developing markers for complex traits can be a much easier thing to accomplish when you have a reference genome to work with: a pre-assembled map of a species’ genes. You can look at the information from your plants and compare it with what is already known. Many crops have reference genomes available.
“Unfortunately, peas don’t,” says LaPlant. So he is drawing his own map and assigning addresses as he goes. It’s working on the frontier of genomic information and LaPlant is breaking trail. “You’d think peas would already have a reference genome because Gregor Mendel was working on peas when he started the field of genetics,” he says. “But no complete version is available to the public yet. Peas have a really large genome which is something like ten times the size of the squash genome.”
Coincidentally, there isn’t a good reference genome for squash, either. We have the right technology; we just need to dedicate the time and resources to doing it. Because once there’s a reference genome and markers for a crop, all of this information can be put into a useable format that breeders can plug into a computer and more efficiently analyze the plants they’re growing.
BANDS AND MARKERS
Over on the other side of the lab, Hultengren is standing in front of a computer. She’s analyzing her plant samples based on markers that LaPlant has developed for squash. “What I’m doing right now is looking to see which gene copies each individual plant has for resistance,” she says.
Hultengren is looking for squash that have two copies of the resistance gene. Resistance to powdery mildew is actually a dominant trait. If a plant has just one gene with the trait, that individual will be resistant in it’s own lifetime but could pass on the recessive susceptibility to it’s offspring. “We don’t want any of these squash to be a carrier for susceptibility.”
To find what she’s looking for, Hultengren uses a technique called electrophoresis. She starts by combining each sample of plant material with a dye, then puts a drop of each sample into a gel. When electricity is run from one end of the gel to the other, it pulls the genetic material along with it, stretching it out into bands of different densities. When luminesced with an ultraviolet lamp, the bands in each sample become visible as a result of the dye.
They appear as a series of columns on Hultengren’s computer.
On the far left side of the screen is a column of very distinct bands: the ladder. It’s a generic standard showing how different densities of genetic material will typically look in a gel. Next to the ladder is another column of less distinct bands: the positive control. It shows how the bands should look for a specific trait in a specific organism. In this case, the positive control shows what powdery mildew resistance looks like in squash.
Across the rest of the screen are more columns, each representing the genetics of a single plant out in the greenhouse. Hultengren is comparing these columns to the positive control to find out which squash she will be saving for the field this coming spring.
As powerful as all these tools are, genetics are only part of a very complex story when it comes to growing food. Because there is absolutely no predicting how all of the genetic interplay within a squash – or any organism – will turn out once the plant starts bumping up against the real world.
“This process doesn’t tell us anything about the actual quality of the squash that these plants might produce,” says Mazourek. A disease-resistant squash is not necessarily a delicious squash. And even if we could identify the plants by genotype for sweetness, how they are raised in the field and handled after the harvest will impact how the squash ultimately tastes on your plate.
Here’s something else to consider: pathogens, like powder mildew, continue to evolve. They’re dynamic, living organisms that constantly strive to overcome any barriers they encounter. But the plants are evolving too, in response to the pathogens and countless other environmental pressures. It’s precisely this sort of real world give-and-take that the marker assisted approach capitalizes on.
“We’re keeping up with diseases and other pressures in real time,” Mazourek says. “This is the advantage that this kind of plant breeding will always have.”
For Mazourek and his team, it’s not about finding a simplistic solution and moving on to the next project as quickly as possible. It’s about continuing to guide the natural evolutionary process of crops so they can excel in a world of ever-changing weather, soil, pests, pathogens, and consumer tastes.
Breeding the perfect crop isn’t easy. Because there is no “perfect” crop. What “works best” is relevant only to a specific set of conditions found at a particular time and place. Change any one of the relevant conditions – change the location or generation – and traits deemed advantageous can suddenly mean nothing.
Our tools may be way more advanced than when we first started sharpening sticks and digging into the earth, but the challenge of farming remains exactly the same: there are no guarantees. We can feel pretty confident that the modern crops we’re breeding are going to keep us fed, but in the end it’s all about probabilities.
What traits are going to give crops the greatest potential for abundance? What technologies and practices are going to make it most likely that farmers will succeed in feeding our communities both today and tomorrow?